EP0248829A1

EP0248829A1 - Control device for a melt electrode.

Info

Publication number: EP0248829A1
Application number: EP86906785A
Authority: EP
Inventors: Otto Stenzel; Friedrich-Werner Thomas
Original assignee: Leybold Heraeus GmbH
Current assignee: Balzers und Leybold Deutschland Holding AG
Priority date: 1985-12-13
Filing date: 1986-11-29
Publication date: 1987-12-16
Anticipated expiration: 2006-11-29
Also published as: JPS63502785A; US4797897A; EP0248829B1; DE3544005C2; DE3544005A1; WO1987003772A1; DE3668748D1

Abstract

Un dispositif sert à régler l'écart entre une électrode de fusion (62) et la surface de matériau fondu (61) dans un four à chauffage direct par arc sous vide. On part du fait que les courts-circuits provoqués par des gouttes de matériau fondu entre l'électrode de fusion (62) et la surface du matériau fondu (61) donnent une indication sur l'état de l'électrode. La fréquence moyenne d'égouttement est à peu près une fonction linéaire de la longueur de l'arc électrique, c'est-à-dire le taux d'égouttement est en rapport hyperbolique avec la longueur de l'arc électrique. Afin d'obtenir dans ces conditions un réglage rapide, on applique à un dispositif de réglage (84) de la fréquence d'égouttement un signal qui correspond à la différence entre la valeur inverse d'un signal de taux d'égouttement formé par un générateur (80) de valeurs moyennes et une valeur de consigne fréquence d'égouttement.A device is used to adjust the gap between a melting electrode (62) and the surface of molten material (61) in a direct vacuum arc furnace. It is assumed that the short circuits caused by drops of molten material between the melting electrode (62) and the surface of the molten material (61) give an indication of the state of the electrode. The average drip frequency is roughly a linear function of the length of the arc, i.e. the rate of dripping is hyperbolic to the length of the arc. In order to obtain a rapid adjustment under these conditions, a signal is applied to a device (84) for adjusting the drip frequency which corresponds to the difference between the inverse value of a drip rate signal formed by a generator (80) of mean values and a drip frequency setpoint.

Description

Device for regulating a melting electrode

The invention relates to a device for regulating the distance of a melting electrode from the surface of the melting material in a vacuum arc furnace.

Various methods are known for producing high-quality metals and metal alloys which have as few foreign body inclusions as possible and which are homogeneous in their structure. One of the best known methods is arc melting, in which an electrode extends to a crucible and by applying an electrical voltage between the electrode and the crucible, the electrode melts at its tip and falls into the crucible as a liquid material. As a rule, the so-called melting electrode is placed on one pole of a DC voltage and the crucible on the other pole of this DC voltage. However, alternations with AC voltage are also possible in order to achieve certain effects.

A major problem with the operation of arc melting furnaces of the type mentioned above is to control the arc length, ie the distance between the lower end of the electrode and the surface of the melting material already in the crucible. If the arc is too long, the electrode and / or the melting material can be heated incorrectly, so that the quality of the melting material is greatly reduced. Since on the one hand the level of the melting material in the crucible is constantly increasing and on the other hand the distance between the end of the electrode and the surface of the crucible cannot be observed directly during operation, special measures must be taken to regulate this distance.

In the case of arc melting furnaces that operate at atmospheric pressure or only slightly below it, the arc length is regulated by maintaining a predetermined arc voltage. At atmospheric pressure, the plasma arc is characterized in that it has a certain voltage gradient, e.g. B. 20 volts per 2.5 cm. The voltage drops on the cathode and on the anode surface add up to an additional 20 volts, so that if you have an arc length of e.g. B. wants to maintain 1.25 cm, brings the electrode into such a position that the arc voltage is 30 V. This can easily be achieved by means of conventional devices which measure and regulate the arc tension.

However, the method described above cannot always be applied to arc melting furnaces which operate in a vacuum. Such arc melting furnaces are used in particular for melting the so-called refractory active metals such as titanium or zirconium and for the production of stainless steels and high-temperature alloys. As the gas pressure surrounding the arc decreases, the voltage gradient of the arc plasma also decreases, and at very low pressures the voltage gradient of the arc plasma can be, for example, only one volt per 2.5 cm. Since the anode and cathode voltage drops z. B. in steel at about 20 volts, the voltage drop on the arc is very small compared to the other voltage drops. Changes in gas content and alloy composition affect the anode and cathode voltage drops on the order of magnitude Voltage drop in the "arc column". As a result, the process of regulating the arc by keeping the arc voltage constant, especially in the case of steels, is fraught with major errors; ie the actual length of the arc will usually deviate greatly from the desired length.

However, an arc melting furnace is already known which makes use of the knowledge to regulate the distance between the electrode and the surface of the material to be melted, that the voltage briefly breaks down even during normal operation at certain time intervals (US Pat. No. 2,942,045). This effect is caused by short circuits caused by liquid metal drops that drip from the electrode into the crucible and briefly connect the electrode to the melting material in the crucible. As long as the duration and frequency of these short circuits are not very long, the arc operates at almost full power, so that there is no significant influence on the heating of the melting material. If the arc becomes shorter, the frequency of the arc short-circuits increases.

According to the known arc melting furnace, the electrode spacing is regulated in such a way that the frequency of the arc short circuits is kept within a certain range. Here, for example, a voltmeter is observed and the time intervals between the individual voltage drops are measured with a stopwatch so that the voltage drops per unit of time can be determined.

Another known device for controlling the electrode spacing in an arc melting furnace is based on the knowledge that voltage fluctuations in the form of positively increasing pulses are superimposed on the arc voltage, each of which occurs for a short period of time, for example 40 milliseconds, at a frequency of 30 Hz (DE -PS 1 212 651). These voltage pulses, the cause of which is not known, are used to regulate the electrode spacing, the voltage curve being divided into a basic component and a two- component is split. The pulse-shaped fluctuations occurring in the second component as voltage, current or impedance fluctuations are detected here, and the electrode spacing is regulated as a function of the repetition frequency of these fluctuations. The overvoltages are counted per unit of time and the electrode is lowered if the number of pulses is too low.

Another known device for arc melting is based on oscilloscopic or graphical observations, which show that during the melting of metals in vacuum, short-term short circuits of 0.1 to 0.3 seconds occur between the electrode and the molten metal surface of the crucible. In addition, it is taken into account that changes in the arc voltage occur, which result from impurities, which in turn are based on changes in the composition or in the pressure of an inactive gas atmosphere or are caused by the deflection of an arc from the electrode to the crucible, these latter voltage changes being smaller are the voltage changes that occur when drops of molten metal bond the electrode to the molten metal ribbon (U.S. Patent 2,915,572). This known device has a device with which the electrode is moved towards the surface of the metal in the crucible by an amount which is at least equal to the difference between the melting rate of the electrode and the rate of increase of the metal surface. The device also has means which are activated due to molten drops between the electrode and the metal surface in a predetermined position of the electrode with respect to the metal surface, in order to move the electrode away from the metal surface by a certain distance. The voltage short circuits are detected by a relay that controls a timepiece.

Furthermore, a device for regulating the electrode spacing is known, in which the drop short circuits between the electrode and the liquid metal surface of the crucible is used as a control criterion (US-PS 45 78 795). The drop short circuits and the associated voltage reductions appear as repeating pulses that correlate closely with the electrode spacing. The number of drop short circuits is added up and each time the number of short circuits has reached a predetermined value, the mean period between the short circuits, based on this predetermined value, and the time in which this value is reached are calculated. A microprocessor is used for this calculation and for displaying the duration of each short circuit. Immediately after the pulse formation of the natural drop short circuits, a computer is used. The standardized impulses are fed to an event register, whereby the quantity of impulses to be counted in can be entered beforehand via a computer control panel and changed on a case-by-case basis. If the content of the event register falls below the specified number of pulses, which is determined by coincidence, a command is sent to a time measuring device and the elapsed time between the respective coincidences is read out. This value serves as a measure of the distance between the electrode and the liquid metal surface. The measured value is renewed each time the specified number of drop short circuits (approx. 100 short circuits) is reached. Such a device has two disadvantages, namely that the measured value is only refreshed after relatively long periods of time, that is to say is not up-to-date, and that if little or no drops occur, the time until a control intervention becomes very long. With a high number of drops, on the other hand, the control intervention is very quick. The determination times for the number of drops represent a dead time. This dead time is of different lengths for different operating states. The timing behavior of the measuring element is non-linear. The phase shift of the signal is therefore dependent on the current operating state. The size of the control intervention in the event of a deviation from the target value must be severely restricted in order to avoid oscillation. A small loop gain must therefore be selected. This results in sluggish disturbance control with large deviations from the setpoint. Finally, yet another method and device for controlling the electrode drive speed in an arc furnace are known, in which the time between two successive drop short circuits is measured and the mean time between a predetermined number of previous short circuits is calculated (US Pat. No. 4,303,797). For example, the time interval between the last ten short circuits is calculated and fed to a controller as the actual value. However, the transient behavior of this device when there are signal changes is disadvantageous.

The invention is therefore based on the object of providing a device with which it is possible to achieve improved electrode regulation.

This object is achieved in that the controller is supplied with a signal which corresponds to the difference between the reciprocal of the drop rate signal formed by the averager and a target time which is equal to the desired mean time between two successive drop short circuits.

The advantage achieved by the invention is, in particular, that the electrode controller can not only be set very sharply, but that it is also possible to achieve very fast regulation with only slight deviations from the setpoint. In this way, the invention optimizes the closed control loop, which has the disadvantage over an open control loop that a control deviation must first occur before the controller can make a correction at all via a manipulated variable. In general control engineering, this disadvantage is mostly attempted to be eliminated by a so-called control with a disturbance variable surcharge. A disturbance variable is measured and fed to the actuator via an auxiliary controller. If a disturbance variable occurs, a correction signal is generated immediately without waiting for a control deviation. In terms of the disturbance variable, this branch represents an open chain, i.e. a controller with all its disadvantages. Below a drop rate, the number of drop short circuits within understood a certain time, while the drop sequence means the time between two drop short circuits. Although the drop rate and the drop sequence are reciprocal values to each other, it is not irrelevant in the present case whether an averaging is carried out first and then a reciprocal value formation or vice versa, because the averaging is an addition process, while the reciprocal formation is a multiplication process. Consequently, the arithmetic mean of a sum of times is not equal to the reciprocal of an arithmetic mean of a sum of rates, ie the sum of the reciprocal of times. This difference has an effect in terms of control technology in that the transition behavior in the case of signal changes, ie the so-called transient behavior, is decisively improved in the solution according to the invention. The dead time is always the same length in the device according to the invention.

An embodiment of the invention is shown in the drawing and will be described in more detail below. Show it:

Fig. 1 shows the voltage curve or the current curve for a typical

Drop short circuit;

2 shows a known control circuit for feeding an electrode in a melting pot;

3 shows a circuit arrangement according to the invention for the drop short-circuit control in a melting electrode connection in analog technology;

FIG. 4 shows a circuit arrangement according to the invention corresponding to FIG. 4, but in digital technology;

5a is a graphical representation which shows the functional relationship between gap width and drop sequence;

5b is a graphical representation showing the functional dependency between gap width and drop frequency.

FIG. 1 shows how current and voltage change, which are present between a melting electrode and a crucible or flow through this distance. It can be seen here that in the case of a direct current on which an alternating current with a small amplitude is superimposed, the voltage U generally remains constant and has a minimum only in points A, B and a maximum in points C, D. If the voltage U has a minimum, then the current has a maximum, see points E and F.

The voltage dips in points A and B each indicate a short circuit, which is caused by a drop of liquid metal, which briefly connects the electrode to the surface of the molten metal in the crucible.

2 shows a known circuit arrangement which is suitable for droplet short-circuit control. On the basis of this known circuit arrangement, the differences and advantages of the invention can be better understood. The actual drop short-circuit control (= drop-short control) is designated here with the reference number 20. It contains several components and works as follows: After the electrical isolation of the arc voltage U _L by an isolating converter 21, the voltage drops due to the drop short circuits are filtered out via a differentiator 22. The limit frequency of the differentiator 22 is designed so that the short circuits can be detected correctly. The filtered pulses are formed in the subsequent trigger circuit 23 and converted in a subsequent monostable multivibrator 24 into standard pulses, ie into pulses of constant amplitude and width.

An integrator, more precisely a PT ₁ element 25 with a fixed integration time, forms the mean pulse value from the resulting normalized pulses. The output variable of this integrator 25 represents the short-circuit frequency actual value U _si , which is measured with a potentiometer 27 Short circuit frequency setpoint U _ss compared and supplied as a difference to the short circuit frequency controller 26. The corresponding digital analogue for this would be the counting of pulses into a digital counter with a fixed time base and the evaluation of the pulse counter reading occurring within a fixed time interval. The number of drop short circuits occurring within a specified time is therefore counted. A ring counter could be used as the averager, which determines the average over all the short-circuits recorded in the ring time. The other circuit parts shown in FIG. 2 have nothing to do with the actual short-circuit frequency control, although they are necessary for the overall control. They take into account other influencing variables on the control, because the drop short circuits are only one of several possible control criteria, which is expressed by the summation element 28. In addition to a constant voltage U _const , the output variable of the short-circuit frequency regulator 26 can be switched on via a switch 29, for example. Furthermore, the summation element 28 can additionally be supplied with the output variable of a general arc voltage regulator 31 via a switch 30. The difference between the arc voltage actual value U _Li and an arc voltage setpoint U _LS , which is tapped by a potentiometer 32, is fed to this arc voltage regulator 31 via a logic element 33. The output groove of this summation element 28 is connected to motor controllers via a controllable switch 34. The switch 34 can be controlled via a relay 35, which is triggered by a voltage generated by a gas detector 36. The motor controllers mentioned control two motors M 1 and M 2, which are provided for the differential gear of the electrode. These motor controllers are of the same type and each have a P controller 37 or 37 ', an I controller 36 or 36', a pulse device 39, 39 ', a rectifier 40, 40' and a resistor, the output of the Resistor 41, 41 'is fed back to a logic element 42, 42' which lies between the P controller 37, 37 'and the I controller 38, 38'. Fefner is the output variable of a tachodynamo TD ₁ or TD ₂ fed back to a link 43 or 43 ', the link 43 between the tap a potentiometer 44 and the P controller 37 is connected, while the link ungseiement 43 'between the output of a reversing amplifier 45, which is also at the tap of the potentiometer 44, and the P controller 37' is connected. Another motor M 3 of the differential gear of the electrodes can be switched on via switches 46, 47, 48, switch 48 being controlled by a relay 49, which in turn is controlled by a short-circuit release circuit 50.

FIG. 3 shows a circuit arrangement according to the invention which corresponds in some details to the arrangement according to FIG. 2. The actual droplet control is now done in a different way. To illustrate, in addition to the circuit arrangement mentioned, the melting pot 60 is also shown, in which the melting material 61, for. B. molten metal or a molten metal alloy. Above this melting material 61, a shrinking electrode 62 is arranged, which is attached to a holding rod 63, which projects through an opening into the crucible and is locked there by means of a flange 64. The part of the holding rod 63 protruding from the crucible 60 is provided with a thread 65 which is guided through a drive nut 66. This drive nut 66 is connected to a gear 67, which in turn is coupled to a motor 68, which drives a tachometer generator 69. A speed control device 70 acts on the motor 68 and is in turn acted upon by signals from the tachometer generator 69. There is a vacuum pump system 71 on the crucible 60, which keeps the inside of the crucible 60 at a predetermined low pressure. A power supply 72 is connected between the bottom of the melting pot 60 and the holding rod 63 of the electrode 62 and applies a voltage between the end of the electrode and the surface of the melting material 61. the so-called arc voltage. The actual value of the arc voltage is passed to a direct current transformer 73, the output signal U _{L i of} which is fed to a logic element 74, which also receives the desired arc value U _LS , which is tapped by a potentiometer 75. The difference between the actual and the target value of the arc voltage is fed to a voltage regulator 75, which outputs a control signal to the speed control 70 via a switch 76. The actual drop sequence control is switched to the speed control 70 via another switch 77. It contains a trigger 78, which is connected to the DC generator 73 and which triggers the drop short-circuit pulses coming from there. The pulses emitted by the trigger 78 can still differ in amplitude and / or pulse width and are therefore fed to a standard pulse generator 79, which uses them to form pulses of a uniform shape. The essential characteristic of the pulses coming from the standard pulse generator is therefore only their time interval, ie the pulse train. The mean value of all the pulses within a predetermined time interval is then formed in a newly switched mean value generator 80. The decisive step of the present invention is that a reciprocal value generator 81 is provided which forms the reciprocal value from the output signal of the mean value generator 80. If the averager 80 has determined a total of X drop short-circuit pulses within T seconds, the pulse rate for this period is X / T. In the reciprocal value generator 61, this value is reversed according to the amount, ie 1 / X per T is calculated. This reciprocal value is then fed as an actual value to a logic element 82, to which a setpoint, which is tapped at a potentiometer 83, is simultaneously fed. The difference between the actual value and the target value then reaches a drip sequence regulator 84, which is connected to the speed control 70 via the switch 77 already mentioned.

As a comparison of FIGS. 2 and 3 shows, the arrangement according to FIG. 2 results in a non-linear circular gain. Measurements carried out (cf., for example, the measurement curves according to the above-mentioned US Pat. No. 4,578,795) show that the mean drop frequency, ie the mean time between two drops, is approximately a linear function of the arc length. Since the reciprocal of the drop rate is the drop sequence, the drop rate has a hyperbolic relationship slope with the arc length (see Fig. 5b). If a controller is now used, or the droplet frequency (rate) is used as the control variable - as shown in FIG. 2, a control loop with a non-constant loop gain is obtained when the system is included. In order to avoid oscillation, the size of the control intervention must be severely restricted if there is a deviation from the setpoint. A small loop gain must therefore be selected. This means sluggish disturbance control with large deviations from the setpoint.

If, on the other hand, as in the arrangement according to FIG. 3, the reciprocal value is formed from the output signal of the mean value generator 80, the deviation signal is proportional to the distance deviation. The time constant and the loop gain are constant, i. H. the controller can be optimally adjusted.

FIG. 4 shows a digital version of the arrangement shown in FIG. 3, the upper region being omitted. It can be seen here that only a pre-memory 90 is connected between the standard pulse generator 79 and averager 80 and a digital-to-analog converter 11 to the output of the drop sequence controller 64. The pre-memory 90 sums up all the pulses that accumulate during a cycle time of the digital evaluation device. It reads out its memory content to the mean value generator 80 at the respective end of the cycle time.

Claims

Patent claims

1.Device for regulating the distance of a melting electrode to the surface of the melting material in a vacuum arc furnace, the short circuits resulting from drops between the melting electrode and the surface of the melting material being used as a control criterion and the short circuits occurring within a predetermined period of time, the so-called drop rate, being determined and an averager, which is connected to a controller which controls an electric drive for the fusing electrode, characterized in that the controller (64) is acted upon by a signal which is the difference between the reciprocal of that formed by the averager (80) Drop rate signal and a target time, which is equal to the target average time between two successive drop shots.

2. Device according to claim 1, characterized in that the mean value generator (60) is preceded by a standard pulse generator (79) which is controlled by a trigger (76) which detects the drop short-circuit signals.

3. Device according to claim 1, characterized in that said controller (84) via a switch (77) with a speed control device (70) can be connected, and that this speed control device (70) simultaneously via a further switch (76) with a voltage regulator (75) is connectable for the arc voltage.

4. Device according to claim 3, characterized in that the speed control device (70) controls a motor (68) which drives a transmission (67) which is connected to a holder (63) of the electrode (62).

5. Device according to claim 2, characterized in that a monostable flip-flop (39, 39 ') is provided as the standard pulse generator (79).

6. Device according to claim 1, characterized in that a PT ₁ element is provided as the mean value generator.

7. Device according to claim 1, characterized in that a ring counter is provided as the averager, which determines the average over all 'detected in the ring time short circuits.

8. The device according to claim 1, characterized in that the mean value generator (80), the reciprocal value generator (81), the target value specification (83) and the above-mentioned controller (64) work digitally and then the output variables in a digital-to-analog converter in one analog size for controlling the feed motor is converted.

9. Device according to claim 1 and 8, characterized in that the signal processing hybrid, d. H. partly digital, partly analog.

10. Device according to claim 6 and 9, characterized in that a pre-memory (90) is set in front of the averager (80), which adds up all the pulses during a cycle time (z. B. 100 ms) of the digital evaluation device and its memory content to the Reads out mean value generator (80) at the end of the cycle time.